Traditionally, clinical measurements (especially, solid phase binding assays) for analytes in blood have been carried out in serum or plasma samples derived from the blood. Partly for this reason, clinicians present the levels of most blood markers in terms of the concentration of the marker in the liquid fraction of the blood sample (e.g., the concentration in plasma or serum fractions derived from the blood sample). Assays carried out in whole blood have had limited acceptance, in part because conventional assays give different signals for a whole blood sample relative to a plasma or serum fraction derived from the same sample. This difference is primarily due to the difference in concentration of the analyte in the different sample types: the concentration of an analyte in a whole blood sample being effectively diluted relative to the concentration of the analyte in the liquid fraction because of the volume occupied by the red blood cells. The difference in signals makes it difficult to compare results to well established reference ranges that have been expressed in terms of the concentration of the analyte in serum or plasma. In addition, the presence of a large volume fraction of cells in a whole blood sample may also interfere with many assay technologies and make it difficult to carry out precise and accurate measurements.
Piironen T, et. al., (2001) Clinical Chem., 74 (4): 703-711 and Tarkkinen, P., et al., (2002) Clinical Chem., 48 (2) 269-277) disclosed, respectively, immunoassays of whole blood samples for prostate specific antigen and C-reactive protein (CRP). Both references report correlation curves comparing the measured levels of the analytes in whole blood relative to the measured levels in the corresponding plasma or serum fractions. In both cases, the authors reported that the graphs could be fit by lines with slopes of roughly 0.56-0.57, the difference from unity being attributed to the average hematocrit (the relative volume of blood occupied by erythrocytes) of the whole blood samples. The authors propose correcting measurements in whole blood samples using the assumption that the samples have a hematocrit value equal to the average hematocrit for the patient population. Because of the wide range for hematocrit values that may be observed in patient samples (the normal range is from 41-50% for adult males and 35-46% for adult females, but hematocrit levels of less than 20% can be observed in severe cases of anemia), this approach could lead to large errors in the reported values.
U.S. Pat. No. 6,475,372 to Ohara T J et al.; U.S. Pat. No. 6,106,778 to Oku N. et al. disclose apparatuses for measuring both the concentration of an analyte in a whole blood sample and the hematocrit of the sample. The measured concentration value is converted to the concentration of analyte in the liquid fraction by using the measured hematocrit value to apply a hematocrit correction. This approach adds complexity to the assay apparatus and may be subject to increased imprecision because the error is a function of the variability in both the assay signal and hematocrit determination.
Other approaches developed for conducting assays on whole blood samples have used instrumentation that employ integrated filters (including lateral flow membranes) or centrifuges to provide for separation of the red blood cells from plasma or serum fractions prior to the measurement of the analyte. While these approaches avoid the need for a hematocrit correction, they add significant cost and complexity to the instrumentation. In addition, filtration has the disadvantages of possible loss of analyte to the filter material and limitations in the sample volume that can be easily processed.